Asl State Types

Pass State

Passes input to output, optionally transforming it with JSONata. Useful for injecting or transforming data. Without Output, the Pass state copies input to output unchanged.

Task State

Executes work via AWS service integrations, activities, or HTTP APIs. Reference service-integrations.md for full details.

Required Fields

• Resource: ARN identifying the task to execute

Optional Fields

• Arguments: Input to the task (replaces JSONPath Parameters)
• Output: Transform the result
• Assign: Store variables from input or result
• TimeoutSeconds: Max task duration (default 99999999, accepts JSONata expression)
• HeartbeatSeconds: Heartbeat interval (must be < TimeoutSeconds)
• Retry: Retry policy array
• Catch: Error handler array
• Credentials: Cross-account role assumption

Choice State

Uses Choices and Condition fields with JSONata boolean expressions to implement branching logic.

Key points:

• Condition must evaluate to a boolean.
• Each Choice Rule can have its own Assign and Output.
• If a rule matches, its Assign/Output are used (not the state-level ones).
• If no rule matches, the state-level Assign is evaluated and Default is followed.
• Default is optional but recommended — without it, States.NoChoiceMatched is thrown.
• Choice states cannot be terminal (no End field).

JSONata supports rich boolean logic:

Wait State

Delays execution for a specified duration or until a timestamp.

Wait Until Timestamp

Timestamps must conform to RFC3339 (e.g., "2026-03-14T01:59:00Z").

A Wait state must contain exactly one of Seconds or Timestamp.

Succeed State

Terminates the state machine (or a Parallel branch / Map iteration) successfully.

Fail State

Terminates the state machine with an error.

Reference error-handling.md for more information.

Parallel State

Executes multiple branches concurrently. All branches receive the same input.

Key points:

• Arguments provides input to each branch’s StartAt state (optional, defaults to state input).
• Result is an array with one element per branch, in the same order as Branches.
• If any branch fails, the entire Parallel state fails (unless caught).
• States inside branches can only transition to other states within the same branch.
• Branch variables are scoped — branches cannot access each other’s variables.
• Use Output on terminal states within branches to pass data back to the outer scope.

Map State

Iterates over a JSON array or a JSON object, processing each element (potentially in parallel). See assets/nested-map-parallel-structures.asl.json for a complete state machine example combining Map + Parallel states.

Key Map Fields

Map ProcessorConfig

The ItemProcessor can include a ProcessorConfig to control execution mode.

• INLINE (default) — iterations run within the parent execution. Use for most cases.
• DISTRIBUTED — iterations run as child executions. Use for large-scale processing (thousands+ items), items read from S3, or when you need per-iteration execution history.

Map Failure Tolerance

Use ToleratedFailurePercentage (0–100) and/or ToleratedFailureCount to allow partial failures. The Map state fails if either threshold is breached.

ItemReader (Distributed Map only, optional)

Specifies a dataset and its location for a Distributed Map state. Omit when iterating over JSON data from a previous state.

Sub-fields:

• Resource — The S3 API action. Use arn:aws:states:::s3:getObject for single files or arn:aws:states:::s3:listObjectsV2 for listing objects.
• Arguments — JSON object specifying Bucket, Key (for getObject), or Prefix (for listObjectsV2). Values accept JSONata expressions.
• ReaderConfig — Configuration object with the following sub-fields:
  • InputType — Required for most sources. Valid values: CSV, JSON, JSONL, PARQUET, MANIFEST.
  • Transformation — Optional. NONE (default) iterates over metadata from ListObjectsV2. LOAD_AND_FLATTEN reads and processes the actual data objects, eliminating the need for nested Maps.
  • ManifestType — Optional. ATHENA_DATA for Athena UNLOAD manifests, S3_INVENTORY for S3 inventory reports. When S3_INVENTORY, do not specify InputType.
  • CSVDelimiter — Optional, for CSV/MANIFEST. Valid values: COMMA (default), PIPE, SEMICOLON, SPACE, TAB.
  • CSVHeaderLocation — Optional, for CSV/MANIFEST. FIRST_ROW uses the file’s first line. GIVEN requires a CSVHeaders array in the config.
  • CSVHeaders — Array of column name strings. Required when CSVHeaderLocation is GIVEN.
  • ItemsPointer — Optional, for JSON files. Uses JSONPointer syntax (e.g., /data/items) to select a nested array or object within the file.
  • MaxItems — Optional. Limits the number of items processed. Accepts an integer or a JSONata expression evaluating to a positive integer. Maximum: 100,000,000.

S3 buckets must be in the same AWS account and Region as the state machine.

ItemSelector (optional)

Overrides the values of input items before they are passed to each iteration. Accepts a JSON object with key-value pairs. Values can be static or JSONata expressions. Can access the map state’s context object. Available in both Inline and Distributed Map states.

ItemBatcher (Distributed Map only, optional)

Groups items into batches for processing. Each child workflow execution receives an Items array and an optional BatchInput object. You must specify at least one of:

• MaxItemsPerBatch — Maximum number of items per batch. Accepts an integer or a JSONata expression evaluating to a positive integer.
• MaxInputBytesPerBatch — Maximum batch size in bytes (up to 256 KiB). Accepts an integer or a JSONata expression evaluating to a positive integer.
• BatchInput — Optional. Fixed JSON merged into each batch. Values accept JSONata expressions.

If both MaxItemsPerBatch and MaxInputBytesPerBatch are specified, Step Functions reduces the item count to stay within the byte limit.

ResultWriter (Distributed Map only, optional)

Controls output formatting and optional export of child workflow execution results to S3. The ResultWriter field cannot be empty — specify at least one of the following combinations:

• WriterConfig only — formats output without exporting to S3.
• Resource + Arguments only — exports to S3 without additional formatting.
• All three — formats and exports.

Sub-fields:

• Resource — arn:aws:states:::s3:putObject
• Arguments — JSON object with Bucket and Prefix. Values accept JSONata expressions.
• WriterConfig — Configuration object:
  • Transformation — NONE (includes execution metadata), COMPACT (output only, preserves array structure), or FLATTEN (output only, flattens nested arrays into one).
  • OutputType — JSON (array) or JSONL (JSON Lines).

When exporting, Step Functions writes SUCCEEDED_n.json, FAILED_n.json, and PENDING_n.json files plus a manifest.json to the specified S3 location. Individual result files are capped at 5 GB. The S3 bucket must be in the same account and Region as the state machine.

Without ResultWriter, the Map state returns an array of child execution results directly. If the output exceeds 256 KiB, the execution fails with States.DataLimitExceeded — use ResultWriter to export to S3 instead.

Distributed Map State Example